Ligand-induced changes in the electrocatalytic activity of atomically precise Au₂₅ nanoclusters

Abstract

Atomically precise gold nanoclusters have shown great promise as model electrocatalysts in pivotal electrocatalytic processes such as the hydrogen evolution reaction (HER) and carbon dioxide reduction reaction (CO₂RR). Although the influence of ligands on the electronic properties of these nanoclusters is well acknowledged, the ligand effects on their electrocatalytic performances have been rarely explored. Herein, using [Au₂₅(SR)₁₈]⁻ nanoclusters as a prototype model, we demonstrated the importance of ligand hydrophilicity versus hydrophobicity in modulating the interface dynamics and electrocatalytic performance. Our first-principles calculations revealed that Au₂₅ protected by hydrophilic –SCH₂COOH ligands exhibits faster kinetics in stripping the thiolate ligand and better HER activity due to enhanced proton transfer facilitated by boosted interface hydrogen bonding. Conversely, Au₂₅ protected by hydrophobic –SCH₂CH₃ ligands demonstrates enhanced CO₂RR performance by minimizing water interference to stabilize the key *COOH intermediate and lower the barrier for CO formation. Experimental validation using synthesized hydrophilic and hydrophobic ligand-protected Au₂₅ nanoclusters (NCs), such as [Au₂₅(MPA)₁₈]⁻ (MPA = mercaptopropionic acid), [Au₂₅(MHA)₁₈]⁻ (MHA = 6-mercaptohexanoic acid), and [Au₂₅(SC₆H₁₃)₁₈]⁻, confirms these findings, where the hydrophilic ligand-protected Au₂₅ NCs exhibit better activity and stability in the HER, while the hydrophobic ligand-protected Au₂₅ NCs achieve higher faradaic efficiency and current density in the CO₂RR. The mechanistic insights in this study provide valuable guidance for the rational design of surface microenvironments in efficient nanocatalysts for sustainable energy applications.

---

1. Introduction

The electrocatalytic reactions such as the hydrogen evolution reaction (HER)^1–4 and carbon dioxide reduction reaction (CO₂RR) have been considered cornerstones in the pursuit of sustainable ways of meeting global energy demands.^5–7 These reactions are critical for the production of clean energy and recycling of carbon, as they enable the conversion of CO₂ into valuable chemicals and fuels and generate hydrogen as a clean energy carrier.^8,9 Given the importance of these processes, the development of highly efficient and stable catalysts for both the CO₂RR and HER is of paramount importance. Atomically precise metal nanoclusters (NCs), renowned for their unique structures and molecule-like properties, have shown exceptional catalytic capabilities due to their quantum confinement effects and a high density of active sites.^10–18 Among these, [Au₂₅(SR)₁₈]⁻ NCs have been particularly recognized for their significant electrocatalytic activities in both CO₂ reduction^12,19–22 and hydrogen evolution.^23–27 The pioneering work by Kauffman et al.²⁸ first highlighted the ability of [Au₂₅(SR)₁₈]⁻ to efficiently reduce CO₂ to CO with high selectivity and low overpotential at commercially viable current densities, establishing the potential of these nanoclusters in electrocatalysis. The bimetallic PtAu₂₄(SR)₁₈ NCs, explored by Kwak et al.,²⁹ demonstrated remarkable catalytic activity for hydrogen production in the HER, surpassing previously reported molecular catalysts and even outperforming traditional platinum catalysts, setting a new benchmark in the field.

In the context of Au₂₅ NCs, the choice and nature of ligands are critical factors that significantly influence their stability and catalytic performance. Ligands act as the outer protective layers, not only preventing aggregation and stabilizing the nanocluster, but also modulating its electronic properties and surface chemistry.^26,30–35 The thiol-based ligands, in particular, are widely used due to their strong S–metal bonds, which provide high stability and a straightforward preparation strategy.^36–38 However, beyond the stabilization effect, the intrinsic properties of the ligands—such as their hydrophilicity or hydrophobicity—can profoundly affect the nanocluster's interaction with reactants and its overall electrocatalytic activity.³⁹ For instance, Kwak et al.²⁵ investigated the impact of different ligands on the HER performance and found that the 3-mercapto-1-propanesulfonic acid (MPS) protected Au₂₅ NC achieved a rate constant of 121 000 s⁻¹ at an overpotential of 0.7 V, which was 11 times higher than that of 1-hexanethiolate (C₆S) protected Au₂₅, highlighting the profound influence of ligand hydrophilicity on catalytic efficiency. Moreover, recent work by Yoo et al.⁴⁰ has shown that the introduction of hydrophobic ligands in a silver nanocluster Ag₂₅ can enhance the CO₂ reduction activity, achieving a faradaic efficiency for CO (FE_CO) of over 90% and a partial current density (j_CO) as high as −240 mA cm⁻² in a gas-fed membrane electrode assembly device. Despite these experimental advancements, atomic-level understanding and elucidation of the specific effects of hydrophilic versus hydrophobic ligands on the electrocatalytic performance of atomically precise NCs have been lacking. This uncertainty underscores the urgent need for a systematic investigation into how the ligand properties influence these catalytic processes. In particular, understanding these interactions is crucial for further optimizing the performance of Au₂₅ NCs and could also provide valuable insights into the rational design of nanocluster-based electrocatalysts for sustainable energy applications.

In this work, we systematically investigated the ligand effect on the interface stability and the electrocatalytic performance of Au₂₅ NCs in both acidic HER and alkaline CO₂RR processes. Utilizing hydrophilic [Au₂₅(SCH₂COOH)₁₇]⁻ and hydrophobic [Au₂₅(SCH₂CH₃)₁₇]⁻ NCs as theoretical models, we first explored how the ligand properties influence the Au–S interface dynamics, the electronic structures, and the electrocatalytic reaction kinetics via the constant potential calculations and ab initio molecular dynamics (AIMD) simulations. The results revealed that the Au–S interface is unstable at the applied electrochemical reduction potential, and hydrophilic [Au₂₅(SCH₂COOH)₁₇]⁻ exhibits faster kinetics for the stripping of the –SR ligand. Moreover, in the acidic environment, hydrophilic Au₂₅ NCs exhibit superior HER performance compared to the hydrophobic one due to the enhanced proton transfer and hydrogen evolution facilitated by the hydrophilic environment. Conversely, under alkaline conditions, hydrophobic Au₂₅ NCs show better CO₂RR activity by promoting the adsorption and stabilization of CO₂ intermediates while minimizing the water interference around the reaction interface. Our theoretical predictions are then validated through the experimental studies, where we synthesized mercaptopropionic acid-protected [Au₂₅(MPA)₁₈]⁻, 6-mercaptohexanoic acid-protected [Au₂₅(MHA)₁₈]⁻, and hexanethiol-protected [Au₂₅(SC₆H₁₃)₁₈]⁻ NCs as model systems and performed the electrochemical tests. These insights provide a deeper understanding of how the ligand environments affect the electrocatalytic activities of Au₂₅ nanoclusters and offer valuable guidance for the rational design of promising nanocatalysts for electrocatalytic applications.

2. Results and discussion

Ligand detachment dynamics

The removal of surface ligands from nanoclusters is a critical step in enhancing their electrocatalytic activity. Fully ligand-protected nanoclusters, such as Au₂₅(SR)₁₈, are typically considered electrochemically inactive and exhibit significant overpotential during the electrocatalytic process.^20,41–44 This inactivity is primarily caused by the passivating effect of the ligands, which block the access of reactants to the metal core and hinder the efficient mass and charge transfer between the nanocluster and the reactants.^45,46 Consequently, the selective removal of ligands from specific sites on the nanocluster surface is necessary to expose the active sites. Our recent studies have provided valuable insights into the mechanism of partial ligand removal in Au₂₅(SCH₃)₁₈ NCs, where the applied reduction potential would induce the spontaneous desulfurization process via breaking the Au–S bonds regardless of the electrolyte environment (acidic, neutral or basic conditions).⁴⁷

Building on these prior findings, we first combined the constant potential calculations and AIMD simulations to explore how the presence of surface hydrophilic and hydrophobic ligands affects the stability of the Au–S interface under acidic electrochemical conditions. We modeled Au₂₅ NCs with ligands aligned along the z-axis, simulating the system in both explicit water slabs and an implicit solvation environment to accurately capture the realistic electrochemical conditions. Our theoretical model includes hydrophilic [Au₂₅(CH₂COOH)₁₈]⁻ and hydrophobic [Au₂₅(CH₂CH₃)₁₈]⁻ NCs (illustrated in Fig. S2†) placed in the simulation box filled with bulk water at an average density of approximately 1 g cm⁻³, which comprises 191 H₂O molecules and one H₃O⁺ ion to simulate the acidic environment.⁴⁸ To explicitly consider the electrode potential, we manually adjusted the number of extra electrons to control the applied potential U (more details on the constant potential calculations can be found in the ESI†). Our AIMD simulations at room temperature (300 K) revealed that the –SR ligands on both the [Au₂₅(CH₂COOH)₁₈]⁻ and [Au₂₅(CH₂CH₃)₁₈]⁻ NCs become unstable when subjected to the applied potential. In the dynamic process, the proton from the solvated H₃O⁺ ion is attracted and adsorbed onto the sulfur atom, weakening the Au–S bonds. When the potential becomes sufficiently negative, this weakening leads to the complete breakage of the Au–S bonds and the formation of the HSR molecule, as illustrated in the AIMD snapshots (Fig. 1, left). Specifically, in the case of [Au₂₅(CH₂CH₃)₁₈]⁻ NCs, the potential (U_RHE = −1.36 V) was not negative enough to facilitate proton adsorption onto the sulfur atom, leading instead to proton transfer and diffusion into the solvent (details provided in Fig. S3†). In contrast, when a lower potential (U_RHE = −1.54 V) was applied, the Au–S bonds in [Au₂₅(CH₂CH₃)₁₈]⁻ NCs were completely broken (Fig. 1b).

Fig. 1 Relative distances between representative atoms (marked with numbers) during the equilibrated AIMD simulations of (a) [Au₂₅(SCH₂COOH)₁₈]⁻ and (b) [Au₂₅(SCH₂CH₃)₁₈]⁻ NCs at around 300 K and at an applied potential of −0.98 V for [Au₂₅(SCH₂COOH)₁₈]⁻ and −1.54 V for [Au₂₅(SCH₂CH₃)₁₈]⁻ in an acidic environment (pH = 0). The corresponding AIMD snapshots after 10 ps are shown on the left, while the insets (right) illustrate the local structures, highlighting the breaking of the Au(surface)–S and subsequent Au(staple)–S bonds. The water molecules in the water layer are shown in line mode. Color codes: yellow, Au; blue, S; grey, C; white, H from the ligand; green, the adsorbed H from the proton in solution. The same color scheme is used in the figures below.

Fig. 1a illustrates the dynamic behavior of [Au₂₅(SCH₂COOH)₁₈]⁻ NC at an applied potential of U_RHE = −0.98 V at 300 K in an acidic environment (pH = 0). The Au¹(surface)–S²⁶ bond first rapidly breaks at around 0.12 ps with the spontaneous proton adsorption at the S site. The Au⁸(staple)–S²⁶ bond then breaks at 4.54 ps, which oscillates, reattaches, and completely breaks again at 7.25 ps, leading to bond dissociation thereafter. This eventually results in the detachment of two Au–S bonds, and in the meantime, the –SR ligand combines with a proton to desorb and dissolve into the solution as a HSCH₂COOH molecule. A similar –SR detachment process occurs for [Au₂₅(SCH₂CH₃)₁₈]⁻ at U_RHE = −1.54 V (Fig. 1b). The proton attack is accompanied by the rapid breaking of the Au⁴(surface)–S²⁸ bond at around 0.18 ps. Afterwards, the Au³(staple)–S²⁸ bond breaks at 5.01 ps and again at 8.37 ps, after which the ligand with adsorbed H dissolves into the solution as a free HSCH₂CH₃ molecule. Interestingly, it seems that the hydrophilic [Au₂₅(SCH₂COOH)₁₈]⁻ NC exhibits faster etching dynamics in the process of ligand stripping. The faster dynamic process is further supported by the higher number of hydrogen bonds observed in the hydrophilic ligand-protected [Au₂₅(SCH₂COOH)₁₈]⁻ system. Specifically, about 113 hydrogen bonds were formed with the surrounding 192 H₂O molecules, as shown in Fig. S4a,† compared to the 80 hydrogen bonds formed in the hydrophobic [Au₂₅(SCH₂CH₃)₁₈]⁻ system. The increased hydrogen bonding in the hydrophilic system facilitates proton transfer to the sulfur atom, thus enhancing the kinetics of ligand detachment. To further investigate whether the carbon chain length of the ligands affects the interaction between the cluster and water molecules, we additionally conducted 10 ps dynamics simulations for Au₂₅ clusters protected by longer hydrophilic and hydrophobic ligands, [Au₂₅(MHA)₁₈]⁻ (MHA = 6-mercaptohexanoic acid) and [Au₂₅(SC₆H₁₃)₁₈]⁻, in the same water environment. The results showed that as the carbon chain length increases, the number of hydrogen bonds between the cluster and water molecules also increases. In the Au₂₅(MHA)₁₈ system, about 276 hydrogen bonds were formed, compared to 245 hydrogen bonds in Au₂₅(SC₆H₁₃)₁₈ (Fig. S4b†). Notably, the difference of 31 hydrogen bonds between the hydrophilic and hydrophobic [Au₂₅(MHA)₁₈]⁻ and [Au₂₅(SC₆H₁₃)₁₈]⁻ is nearly identical to the 33 hydrogen bond difference observed in [Au₂₅(SCH₂COOH)₁₈]⁻ and [Au₂₅(SCH₂CH₃)₁₈]⁻. This similarity indicates that, although the total number of hydrogen bonds increases with longer carbon chains, the differences in hydrogen bond counts between Au₂₅ NCs with ligands of similar chain lengths are primarily attributed to the ligand hydrophilicity. Therefore, the accuracy of the simplified ligand calculations can be reliably ensured. These observations collectively underscore the critical role of applied potential, ligand environment, and hydrophilicity in driving the detachment of –SR ligands from the nanocluster surface—a necessary step for exposing active sites and enhancing the electrocatalytic performance of metal NCs.

HER performance in an acidic environment

Building on our above investigation into the dynamics of potential-induced ligand detachment, we now examine the electrocatalytic performance of the two dethiolated [Au₂₅(SR)₁₇]⁻ structures, where one –SR ligand has been removed. To compare the HER performance of dethiolated [Au₂₅(SCH₂COOH)₁₇]⁻ and [Au₂₅(SCH₂CH₃)₁₇]⁻, we constructed models that include both the explicit solvation with 50 water molecules and the implicit solvation effects.⁴⁹ The HER testing was conducted under acidic conditions at pH = 0, providing an optimal environment for hydrogen evolution.

Fig. 2a and b present the U–ΔG plots derived from the work function fitting for [Au₂₅(SCH₂COOH)₁₇]⁻ and [Au₂₅(SCH₂CH₃)₁₇]⁻, respectively. These plots were obtained by calculating the N_e–U and U–G relationships (additional details are provided in Fig. S5†). The results indicate that the first step of the HER, the Volmer reaction, is thermodynamically favorable across the entire potential range, as evidenced by ΔG values consistently below zero. This suggests that the adsorption of protons onto the nanocluster surface with exposed Au sites occurs spontaneously. However, the second step of the HER, the Heyrovsky reaction—which involves the formation of H₂—exhibits ΔG values greater than zero when the potential is not sufficiently negative. As the potential becomes more negative, the ΔG values gradually decrease, eventually falling below zero, indicating that the formation of H₂ becomes thermodynamically favorable only at more negative potentials. Therefore, these constant potential thermodynamic calculations lead us to conclude that, for both [Au₂₅(SCH₂COOH)₁₇]⁻ and [Au₂₅(SCH₂CH₃)₁₇]⁻, the formation of H₂ in the second step of the HER (Heyrovsky reaction) is the rate-determining step.

Fig. 2 Free energy changes (ΔG) for HER steps as a function of electrode potential for (a) [Au₂₅(SCH₂COOH)₁₇]⁻ and (b) [Au₂₅(SCH₂CH₃)₁₇]⁻. Energy profiles sampled by slow-growth (SG)-AIMD for the HER rate-determining step of (c) [Au₂₅(SCH₂COOH)₁₇]⁻ and (d) [Au₂₅(SCH₂CH₃)₁₇]⁻at similar potentials (vs. RHE) with representative initial state (IS) and final state (FS) structures.

While thermodynamic calculations provide valuable insights into the feasibility of reaction steps, they often overlook the kinetic barriers that determine the rate at which these reactions proceed. To address this, we employed the slow-growth method within the framework of constrained kinetics to calculate the energy barriers associated with the rate-determining step of the HER for both [Au₂₅(SCH₂COOH)₁₇]⁻ and [Au₂₅(SCH₂CH₃)₁₇]⁻. Fig. 2c and d depict the energy barriers as a function of the constrained variable (CV) (Fig. S1†) at an applied potential of −0.66 V. As the constraint increases, the energy barrier increases until it peaks at the point where two hydrogen atoms—one adsorbed at the Au bridge site and the other from H₃O⁺ in the water environment—combine to form H₂ and then desorb. The analysis of the reaction pathway shows that the [Au₂₅(SCH₂COOH)₁₇]⁻ system reaches this maximum energy barrier earlier, with a free energy of 0.42 eV, indicating a lower energy requirement for the rate-determining step. In contrast, at a similar potential, the energy barrier for the HER rate-determining step in the [Au₂₅(SCH₂CH₃)₁₇]⁻ system is higher, at 0.54 eV. These results suggest that the hydrophilic [Au₂₅(SCH₂COOH)₁₇]⁻ nanocluster exhibits superior HER performance compared to the hydrophobic [Au₂₅(SCH₂COOH)₁₇]⁻ nanocluster. The lower kinetic barrier in the hydrophilic [Au₂₅(SCH₂COOH)₁₇]⁻ system facilitates a more efficient hydrogen evolution reaction. This improved performance can be attributed to the enhanced interaction between the hydrophilic ligands and the surrounding aqueous environment, which promotes the formation and desorption of H₂, thereby making the HER process more favorable in [Au₂₅(SCH₂COOH)₁₇]⁻.

CO₂RR performance in an alkaline environment

We further evaluated the CO₂RR performance of dethiolated [Au₂₅(SCH₂COOH)₁₇]⁻ and [Au₂₅(SCH₂CH₃)₁₇]⁻ by constant potential thermodynamic and kinetic calculations. The CO₂RR performance testing was carried out under alkaline conditions at pH = 14. Note that the alkaline environments are favorable for CO₂ reduction and significantly suppress the HER, minimizing the interference and allowing for a more targeted comparison of CO₂RR activity between the two nanoclusters.

Fig. 3a and b show the U–ΔG plots derived from work function fitting for [Au₂₅(SCH₂COOH)₁₇]⁻ and [Au₂₅(SCH₂CH₃)₁₇]⁻, respectively (additional details are provided in Fig. S6†). These plots detail the ΔG values associated with the four key steps of the CO₂RR. Notably, the reaction of *COOH → *CO + OH⁻ exhibits the highest ΔG value when the potential is higher than −1.0 V, indicating that it is the thermodynamically most challenging step. In contrast, the other three steps show lower ΔG values, suggesting that they are either spontaneous or more easily facilitated under the same potential conditions. Therefore, we conclude that the conversion of *COOH to *CO is the rate-determining step for the CO₂RR in both the nanocluster systems.

Fig. 3 Free energy changes (ΔG) for CO₂RR steps as a function of electrode potential for (a) [Au₂₅(SCH₂COOH)₁₇]⁻ and (b) [Au₂₅(SCH₂CH₃)₁₇]⁻. Energy profiles sampled by SG-AIMD for the rate-determining CO₂RR step of (c) [Au₂₅(SCH₂COOH)₁₇]⁻ and (d) [Au₂₅(SCH₂CH₃)₁₇]⁻ at similar potentials (vs. RHE) with representative IS and FS structures.

Next, we simulated the constant potential kinetics of this rate-determining step for both nanoclusters at similar applied potentials. As depicted in Fig. 3c and d, we tracked the energy barrier as a function of the constrained variable, which corresponds to the C–O bond distance within the *COOH intermediate (Fig. S1†). The energy barrier increases as the CV constraint grows, reaching a peak when *COOH is converted into *CO, and in the meantime, the generated OH⁻ is released into the water environment. The kinetic analysis revealed that the energy barrier for the [Au₂₅(SCH₂CH₃)₁₇]⁻ NC is relatively lower, with a smoother curve and a peak barrier of 0.76 eV. In contrast, the [Au₂₅(SCH₂COOH)₁₇]⁻ NC exhibits a significantly higher energy barrier, with free energy peaking at 1.09 eV. This higher energy input indicates that the CO₂ reduction process in [Au₂₅(SCH₂COOH)₁₇]⁻ is less efficient compared to that in [Au₂₅(SCH₂CH₃)₁₇]⁻. The superior CO₂RR performance of [Au₂₅(SCH₂CH₃)₁₇]⁻ can be attributed to the hydrophobic nature of the –SCH₂CH₃-protected nanocluster. The hydrophobic environment likely facilitates the desorption of the OH⁻ species and the formation of *CO by minimizing interactions with the surrounding water molecules. This reduced interaction lowers the energy barrier for the rate-determining step, making the reaction pathway more favorable and thereby enhancing the overall CO₂ reduction efficiency.

To further substantiate the selectivity of the CO₂RR over the HER under alkaline conditions, we performed the constrained kinetic simulations for the competitive HER process on [Au₂₅(SCH₂COOH)₁₇]⁻ and [Au₂₅(SCH₂CH₃)₁₇]⁻ NCs. The results revealed that while the Volmer reaction (water dissociation and proton adsorption) is feasible (with a barrier less than 0.7 eV) (Fig. S7†), the Heyrovsky reaction (H₂ formation) encounters insurmountable barriers, where the splitting of the second H₂O molecule and subsequent H₂ generation display progressively increasing energy barriers (Fig. S8†). This indicates that the HER is kinetically hindered under alkaline conditions, thereby ensuring high selectivity for the CO₂RR on these nanoclusters.

The catalytic performance difference in the HER and CO₂RR

The optimized local structures of the Au(surface) and Au(staple) sites for both clusters are presented in Fig. 4a and b, while the adsorption configurations of the key *H and *COOH intermediates are provided in Fig. S9.†Fig. 4c presents the radial distribution function (RDF) between the Au active sites in [Au₂₅(SCH₂COOH)₁₇]⁻/[Au₂₅(SCH₂CH₃)₁₇]⁻ and the hydrogen atoms of water molecules during the rate-determining steps of the HER and CO₂RR. The solid lines represent the RDF for both Au(surface) and Au(staple) during the HER, while the dashed lines depict the RDF for Au(staple) during the CO₂RR.

Fig. 4 Local structures of the Au(surface) and Au(staple) sites in (a) [Au₂₅(SCH₂COOH)₁₇]⁻ and (b) [Au₂₅(SCH₂CH₃)₁₇]⁻ NCs. (c) Radial distribution function (RDF) of Au active sites with hydrogen atoms from surrounding H₂O during the kinetic process of the rate-determining steps in the HER (solid lines) and CO₂RR (dashed lines) for [Au₂₅(SCH₂COOH)₁₇]⁻ and [Au₂₅(SCH₂CH₃)₁₇]⁻. (d) Bader charge analysis of Au(surface) and Au(staple) sites in the NC catalysts before and after the adsorption of key *H and *COOH intermediates in acidic HER and alkaline CO₂RR.

In the HER rate-determining step, the RDF considers both Au(staple) and Au(surface) sites because the *H adsorption occurs at both the active Au sites. The first RDF peak corresponds to the closest interactions between the Au atoms and the hydrogen atoms of surrounding water molecules, occurring around 1.7 Å. For [Au₂₅(SCH₂COOH)₁₇]⁻, the g(r) value at this peak is 3.25, whereas for [Au₂₅(SCH₂CH₃)₁₇]⁻, it is significantly lower at 0.98. The higher g(r) value for [Au₂₅(SCH₂COOH)₁₇]⁻ indicates stronger and more frequent interactions between the Au active sites and the hydrogen atoms. This can be attributed to the hydrophilic nature of the –SCH₂COOH ligands, which draw water molecules closer to the Au surface, enhancing proton transfer and interaction with active sites, thereby improving the HER performance. In contrast, the lower g(r) value for [Au₂₅(SCH₂CH₃)₁₇]⁻ reflects weaker interactions due to the hydrophobic nature of the –SCH₂CH₃ ligands, which repel water molecules and limit hydrogen bonding near the Au surface, thereby reducing the HER efficiency. The second RDF peak, occurring around 2.9 Å, reflects the next shell of water molecules interacting with the Au active sites. Here, the g(r) value is 5.66 for [Au₂₅(SCH₂COOH)₁₇]⁻ and 4.08 for [Au₂₅(SCH₂CH₃)₁₇]⁻. The higher and sharper peak for the hydrophilic cluster indicates a more ordered and denser water structure around the Au atoms, which facilitates proton transfer and H₂ formation during the HER, further explaining its superior HER performance. Conversely, the less structured water network around the hydrophobic cluster hinders efficient proton transfer, limiting its HER activity.

For the CO₂RR rate-determining step, the RDF focuses on the interaction between Au(staple) and the hydrogen atoms of water molecules because *CO₂ and *COOH adsorption as well as the conversion to *CO occur exclusively at the Au(staple) site. Interestingly, [Au₂₅(SCH₂COOH)₁₇]⁻ also shows a higher and sharper peak compared to [Au₂₅(SCH₂CH₃)₁₇]⁻, indicating a more ordered and dense water structure around Au(staple). However, this structured water environment may hinder the adsorption and stabilization of CO₂RR intermediates such as *COOH by increasing the energy barrier for *COOH conversion to *CO. In contrast, the less structured water environment in [Au₂₅(SCH₂CH₃)₁₇]⁻ minimizes water interference with CO₂RR intermediates, facilitating *COOH to *CO conversion with a lower barrier and enhancing CO₂RR performance.

Fig. 4d presents the Bader charge analysis of the Au(surface) and Au(staple) sites for both nanoclusters before and after the adsorption of key intermediates (*H and *COOH) in the HER and CO₂RR. These data reveal distinct differences in how the two systems interact with reactants, helping explain their differing catalytic performances. In the * state (without adsorption), both systems show similar Bader charges on the Au(surface) (0.13), but the Au(staple) in [Au₂₅(SCH₂COOH)₁₇]⁻ has a significantly lower charge (0.03) compared to [Au₂₅(SCH₂CH₃)₁₇]⁻ (0.11). The lower charge on Au(staple) in the hydrophilic cluster suggests a lower electron density, which could promote better interaction with protons facilitating proton transfer and enhancing the hydrogen evolution. When hydrogen is adsorbed (*H state), both systems exhibit negative charges on the Au sites, indicating electron transfer from H to Au. However, the Au(staple) site in [Au₂₅(SCH₂CH₃)₁₇]⁻ holds a higher negative charge (−0.04) compared to [Au₂₅(SCH₂COOH)₁₇]⁻ (−0.02), suggesting a stronger electron interaction with hydrogen. This stronger interaction likely hinders H₂ desorption in the hydrophobic cluster, increasing the energy barrier for H₂ release. In contrast, the more balanced electron distribution in the hydrophilic cluster allows for easier H₂ desorption, contributing to its superior HER performance. In the *COOH state, a crucial intermediate for the CO₂RR, the Au(staple) in [Au₂₅(SCH₂CH₃)₁₇]⁻ shows a more negative charge (−0.12) than in [Au₂₅(SCH₂COOH)₁₇]⁻ (−0.08). This indicates that the hydrophobic cluster better stabilizes the *COOH intermediate by receiving more electrons from the adsorbed species. This enhanced stabilization lowers the energy barrier for *COOH reduction to *CO, explaining the superior CO₂RR performance of [Au₂₅(SCH₂CH₃)₁₇]⁻. Conversely, the less negative charge on Au(staple) in the hydrophilic cluster results in weaker stabilization of *COOH, leading to a higher energy barrier for this reduction step and contributing to its lower CO₂RR efficiency.

Fig. 5 illustrates the projected density of states (PDOS) and corresponding d-band centers (ε_d) of active Au sites for [Au₂₅(SCH₂COOH)₁₇]⁻ and [Au₂₅(SCH₂CH₃)₁₇]⁻ NCs, along with their key intermediates involved in the HER and CO₂RR. The ε_d values offer valuable insights into the electronic properties of these catalysts and their interaction with reaction intermediates. In the pristine nanocluster (* state), [Au₂₅(SCH₂COOH)₁₇]⁻ exhibits an ε_d of −3.34 eV, while [Au₂₅(SCH₂CH₃)₁₇]⁻ has a more negative ε_d of −3.61 eV. A more negative ε_d typically indicates a lower energy of the d-band center, which correlates with stronger binding of adsorbates. For the *H state, [Au₂₅(SCH₂COOH)₁₇]⁻ has an ε_d of −3.14 eV, whereas [Au₂₅(SCH₂CH₃)₁₇]⁻ shows a more negative ε_d of −3.38 eV. This stronger interaction in [Au₂₅(SCH₂CH₃)₁₇]⁻ reflects tighter hydrogen binding, which could make H₂ desorption more difficult and increase the energy barrier for the HER. In contrast, the weaker interaction in [Au₂₅(SCH₂COOH)₁₇]⁻ facilitates easier H₂ desorption, contributing to its superior HER activity. In the *COOH state, the ε_d for [Au₂₅(SCH₂COOH)₁₇]⁻ is −3.65 eV, compared to a more negative ε_d of −3.78 eV for [Au₂₅(SCH₂CH₃)₁₇]⁻. The more negative ε_d in [Au₂₅(SCH₂CH₃)₁₇]⁻ indicates stronger stabilization of the *COOH intermediate, lowering the energy barrier for *COOH to *CO conversion and enhancing the CO₂RR performance.

Fig. 5 PDOS and corresponding d-band centers (ε_d) for [Au₂₅(SCH₂COOH)₁₇]⁻ and [Au₂₅(SCH₂CH₃)₁₇]⁻, and their key intermediates. (a) [Au₂₅(SCH₂COOH)₁₇]⁻, (b) *H adsorbed on [Au₂₅(SCH₂COOH)₁₇]⁻, (c) *COOH adsorbed on [Au₂₅(SCH₂COOH)₁₇]⁻, (d) [Au₂₅(SCH₂CH₃)₁₇]⁻, (e) *H adsorbed on [Au₂₅(SCH₂CH₃)₁₇]⁻, and (f) *COOH adsorbed on [Au₂₅(SCH₂CH₃)₁₇]⁻.

In summary, our above theoretical analysis revealed that the catalytic differences between [Au₂₅(SCH₂COOH)₁₇]⁻ and [Au₂₅(SCH₂CH₃)₁₇]⁻ in the HER and CO₂RR are contributed both by their surface interaction with the water environment and their electronic properties. The RDF analysis shows that [Au₂₅(SCH₂COOH)₁₇]⁻ with hydrophilic ligands enhances proton transfer and H₂ desorption, leading to better HER performance. In contrast, the hydrophobic nature of [Au₂₅(SCH₂CH₃)₁₇]⁻ reduces water interaction and lowers the HER efficiency. The Bader charge and d-band center analyses reveal that [Au₂₅(SCH₂CH₃)₁₇]⁻ strongly stabilizes the *COOH intermediate, lowering the energy barrier for *CO formation, thereby excelling in the CO₂RR, while [Au₂₅(SCH₂COOH)₁₇]⁻ shows weaker stabilization of *COOH, leading to higher CO₂RR barriers.

Experimental validation

To validate the computational findings, which indicate that the hydrophilic ligand-protected Au₂₅ cluster exhibits superior HER performance in an acidic environment, whereas the hydrophobic ligand-protected Au₂₅ cluster excels in the CO₂RR in an alkaline environment, we conducted further experimental investigations. Specifically, we synthesized [Au₂₅(MPA)₁₈]⁻ as the hydrophilic model and [Au₂₅(SC₆H₁₃)₁₈]⁻ as its hydrophobic counterpart. To confirm the successful synthesis of [Au₂₅(MPA)₁₈]⁻ and [Au₂₅(SC₆H₁₃)₁₈]⁻ NCs, we conducted the UV-Vis absorption spectroscopy measurements. The UV-Vis spectra of the two Au₂₅ NCs, provided in Fig. S10a and b,† exhibit distinct absorption peaks at 400 nm, 450 nm, and 670 nm, which are consistent with the characteristic optical behavior of Au₂₅ NCs reported in previous studies.^22,25

In the HER experiments conducted in 0.5 M H₂SO₄ (Fig. 6a), [Au₂₅(MPA)₁₈]⁻ demonstrated a lower overpotential of 445 mV at a current density of 10 mA cm⁻² compared to 541 mV for [Au₂₅(SC₆H₁₃)₁₈]⁻, indicating an enhanced hydrogen evolution activity of the [Au₂₅(MPA)₁₈]⁻ NC. Additionally, the Tafel slope for [Au₂₅(MPA)₁₈]⁻ was slightly lower (106 mV dec⁻¹) than that of [Au₂₅(SC₆H₁₃)₁₈]⁻ (111 mV dec⁻¹), suggesting faster reaction kinetics (Fig. 6b). The stability tests over 8.5 hours of continuous operation (Fig. 6c) revealed that [Au₂₅(MPA)₁₈]⁻ maintained a stable current density, whereas [Au₂₅(SC₆H₁₃)₁₈]⁻ exhibited a reduction of about 10% in the current density, indicating that the hydrophilic [Au₂₅(MPA)₁₈]⁻ is more stable under acidic HER conditions.

Fig. 6 Electrocatalytic HER performance in 0.5 M H₂SO₄. (a) LSV curves of [Au₂₅(MPA)₁₈]⁻ and [Au₂₅(SC₆H₁₃)₁₈]⁻. (b) Tafel slopes of the catalysts. (c) Chronoamperometric plots at 10 mA cm⁻² for the HER stability testing.

For the CO₂RR performance evaluation, linear sweep voltammetry (LSV) was performed using a CO₂ flow cell with 1.0 M KOH electrolyte. As shown in Fig. 7a, the current density difference between CO₂-saturated and N₂-saturated electrolytes for [Au₂₅(SC₆H₁₃)₁₈]⁻ was larger than that for [Au₂₅(MPA)₁₈]⁻, indicating that [Au₂₅(SC₆H₁₃)₁₈]⁻ is more favorable for CO₂ reduction to CO. The faradaic efficiency for CO (FE_CO) at various potentials (Fig. 7b) showed that [Au₂₅(SC₆H₁₃)₁₈]⁻ achieved a higher FE_CO, reaching 97.64% at −0.48 V. In terms of the partial current density for CO production (j_CO) (Fig. 7c), [Au₂₅(MPA)₁₈]⁻ exhibited slightly higher values in the potential range of −0.18 V to −0.48 V, whereas [Au₂₅(SC₆H₁₃)₁₈]⁻ outperformed in the lower potential range of −0.58 V to −0.98 V. The Tafel slopes (Fig. S11†) further indicated the faster reaction kinetics for [Au₂₅(SC₆H₁₃)₁₈]⁻ (199 mV dec⁻¹) than that for [Au₂₅(MPA)₁₈]⁻ (242 mV dec⁻¹). Additionally, the turnover frequency (TOF) values (Fig. 7d) showed that [Au₂₅(MPA)₁₈]⁻ has slightly higher TOF between −0.18 V and −0.48 V, while [Au₂₅(SC₆H₁₃)₁₈]⁻ exhibits significantly higher TOF values from −0.58 V to −0.98 V, indicating the much enhanced activity in lower potential ranges. The long-term stability tests over 8.5 hours of continuous operation demonstrated that [Au₂₅(MPA)₁₈]⁻ showed a slight decrease of around 4% in FE_CO and current density (Fig. 7e), whereas [Au₂₅(SC₆H₁₃)₁₈]⁻ maintained the stable FE_CO and current density (Fig. 7f), suggesting that the hydrophobic nanocluster is more stable under alkaline CO₂RR conditions.

Fig. 7 Electrocatalytic CO₂RR performance in a CO₂ flow cell with 1 M KOH electrolyte. (a) LSV curves of [Au₂₅(MPA)₁₈]⁻ and [Au₂₅(SC₆H₁₃)₁₈]⁻ in N₂ and CO₂ environments. (b) Faradaic efficiency for CO (FE_CO) of [Au₂₅(MPA)₁₈]⁻ and [Au₂₅(SC₆H₁₃)₁₈]⁻ at different potentials. (c) Fractional current density of CO at various potentials. (d) Turnover frequency (TOF) of CO at different potentials. (e and f) Long-term stability tests of [Au₂₅(MPA)₁₈]⁻ and [Au₂₅(SC₆H₁₃)₁₈]⁻ for the CO₂RR at −0.48 V.

Recognizing that the difference in the ligand chain length could also influence the electrocatalytic activity, we further synthesized and tested Au₂₅ clusters protected by ligands with similar chain lengths. Specifically, [Au₂₅(MHA)₁₈]⁻ (MHA = 6-mercaptohexanoic acid) was chosen as a hydrophilic analog with a chain length comparable to that of [Au₂₅(SC₆H₁₃)₁₈]⁻. These additional experiments, detailed in the ESI,† provide a comprehensive analysis of how both ligand hydrophilicity and chain length affect the catalytic performance of Au₂₅ NCs. The UV-Vis spectra of the synthesized [Au₂₅(MHA)₁₈]⁻ NCs exhibit distinct absorption peaks at 400 nm, 450 nm, and 670 nm (Fig. S10c†), confirming the successful synthesis of Au₂₅ NCs.^21,22,25

The LSV curves were measured in a flow cell with 1.0 M KOH electrolyte (Fig. S12a†), revealing that [Au₂₅(SC₆H₁₃)₁₈]⁻ NCs are more favorable for CO₂ reduction to CO. In the comparison of FE_CO at various potentials (Fig. S12b†), [Au₂₅(SC₆H₁₃)₁₈]⁻ NCs demonstrate higher CO₂ reduction selectivity than [Au₂₅(MHA)₁₈]⁻ NCs, with the maximum FE_CO for [Au₂₅(MHA)₁₈]⁻ reaching only 88%. As shown in Fig. S12c,† the j_CO of [Au₂₅(SC₆H₁₃)₁₈]⁻ NCs is up to twice as high as that of [Au₂₅(MHA)₁₈]⁻ NCs. The Tafel slopes (Fig. S13†) further highlight the faster reaction kinetics for [Au₂₅(SC₆H₁₃)₁₈]⁻ NCs (199 mV dec⁻¹) compared to [Au₂₅(MHA)₁₈]⁻ (498 mV dec⁻¹). Additionally, the TOF values (Fig. S12d†) indicate that [Au₂₅(SC₆H₁₃)₁₈]⁻ exhibits significantly higher TOF than [Au₂₅(MHA)₁₈]⁻ across various potentials.

These experimental results have well corroborated our computational predictions, confirming that the hydrophilic ligand-protected Au₂₅ NCs are more effective for the HER in acidic environments due to enhanced proton transfer facilitated by the hydrophilic ligands. Conversely, the hydrophobic ligand-protected Au₂₅ NCs demonstrate superior CO₂RR activity under alkaline conditions by promoting adsorption of CO₂RR intermediates and minimizing the water interference, thereby enhancing CO₂ reduction efficiency. Further experiments with ligands of similar chain lengths to the hydrophilic and hydrophobic Au₂₅ NCs reveal that both ligand hydrophilicity and chain length influence catalytic performance. However, the hydrophilicity of the ligands plays a dominant role in determining the overall catalytic behavior. This alignment between computational and experimental findings underscores the critical role of ligand properties in tuning the electrocatalytic performance of atomically precise metal NCs, providing valuable insights for the rational ligand design of nanocatalysts in sustainable energy applications.

3. Conclusion

In this study, we have systematically explored how the ligand hydrophilicity and hydrophobicity influence the interface dynamics and the electrocatalytic performance of Au₂₅ NCs in important reactions such as the HER and CO₂RR. Our computations revealed that the hydrophilic [Au₂₅(SCH₂COOH)₁₇]⁻ NC exhibits faster breakage of the Au–S bond and desorption of the thiolate ligand into solution at the applied reduction potential. The dethiolated [Au₂₅(SCH₂COOH)₁₇]⁻ NC is predicted to exhibit superior acid HER activity due to enhanced proton transfer and efficient hydrogen evolution facilitated by stronger hydrogen bond interactions with water molecules. Conversely, the hydrophobic [Au₂₅(SCH₂CH₃)₁₇]⁻ NC demonstrates enhanced alkaline CO₂RR performance by stabilizing key CO₂RR intermediates and minimizing the water interference to lower the barrier for CO formation. The electrochemical experiments using synthesized [Au₂₅(MPA)₁₈]⁻ and [Au₂₅(SC₆H₁₃)₁₈]⁻ as models, which have similar ligand molecular weights, have further validated our predictions. Hydrophilic [Au₂₅(MPA)₁₈]⁻ showed better activity, lower overpotential, and greater stability in the acidic HER testing, while hydrophobic [Au₂₅(SC₆H₁₃)₁₈]⁻ achieved higher faradaic efficiency, current density, and stability in alkaline CO₂RR. Moreover, the results from clusters with similar chain lengths, such as [Au₂₅(MHA)₁₈]⁻ and [Au₂₅(SC₆H₁₃)₁₈]⁻, further corroborated our computations, highlighting that both ligand hydrophilicity and chain length influence the catalytic performance, with hydrophilicity being the dominant factor in determining the overall catalytic behavior. These results not only provide an atomic level understanding of the ligand effects on the interface dynamics and electrocatalytic performance of metal NCs, but also stimulate ligand engineering as a facile strategy to optimize the catalytic activity of nanocluster catalysts for specific catalytic targets in future electrocatalysis research.

4. Computational methods and experimental section

Ligand removal dynamics

The dynamics of ligand removal from the nanocluster were investigated using the CP2K package (version 2023.1).⁵⁰ The simulations employed the Perdew–Burke–Ernzerhof (PBE) functional within the generalized gradient approximation (GGA) in combination with the hybrid Gaussian/Plane-Wave (GPW) scheme.⁵¹ The computational domain was defined with a periodic lattice size of 23 × 23 × 28 ų, including 192 water molecules to simulate the solvent environment. Molecular dynamics (MD) simulations were performed in the canonical (NVT) ensemble using Nose–Hoover thermostats to maintain a constant temperature of 300 K.^52,53 A time step of 1.0 fs was used throughout the simulations, which were run for a total of 10 ps. Dispersion corrections were applied using the DFT-D3 method to account for van der Waals interactions.^54,55 The electronic structure was described by density functional theory (DFT) with spin-polarization and a mixed double-ζ Gaussian and plane-wave basis set, with an energy cutoff of 400 Ry.⁵⁶

Thermodynamic calculations for the HER and CO₂RR

Spin-polarized DFT calculations were carried out using the Vienna Ab initio Simulation Package (VASP, version 5.4.4).⁵⁷ The exchange–correlation effects were described using the Perdew–Burke–Ernzerhof (PBE) functional within the generalized gradient approximation (GGA).⁵¹ The projector-augmented wave (PAW) method was employed to account for electron–ion interactions with a cutoff energy of 400 eV.⁵⁸ For these calculations, the computational cell was set with a lattice size of 23 × 23 × 28 ų, including 50 water molecules to adapt to the specific requirements of the reaction environments. The empirical DFT-D3 dispersion correction was applied to include van der Waals interactions.⁵⁹ Brillouin zone integration was performed using a gamma centered 1 × 1 × 1 k-point grid, with convergence criteria set at 0.05 eV Å⁻¹ for forces and 10⁻⁴ eV per cell for energy. The electrode potential was determined via tuning the work function (additional details are provided in the ESI†).

Dynamic calculations for the HER and CO₂RR

Ab initio molecular dynamics (AIMD) simulations were conducted using the Nose–Hoover thermostat in the canonical (NVT) ensemble at 300 K with a time step of 1 fs.^52,60 Thermodynamic integrations and the “slow-growth” method were employed to derive the free energy profiles for the rate-determining steps in the HER and CO₂RR. The collective variable (CV) increment was set to 0.0008 Å, and the simulation time was 3 ps. Reaction barriers and reaction energies were determined by integrating the free-energy gradients to generate the free energy profiles.^61,62

Chemicals and materials

Tetrachloroauric(III) acid (HAuCl₄·3H₂O, >99.99%), tetraoctylammonium bromide (TOAB, 98%), 1-hexanethiol (97%), 3-mercaptopropionic acid (MPA, 98%), 6-mercaptohexanoic acid (MHA, 98%), sodium borohydride (NaBH₄), sodium hydroxide (NaOH), potassium hydroxide (KOH), Nafion solution (5 wt%), and solvents: reverse osmosis water, methanol, ethanol, acetonitrile (CH₃CN), dichloromethane (CH₂Cl₂) and tetrahydrofuran (THF) were used. High purity (>99.999%) Ar and CO₂ gases were used.

Preparation of the [Au₂₅(MPA)₁₈]⁻ NCs

In a typical synthesis of [Au₂₅(MPA)₁₈]⁻ clusters, 4 mg of HAuCl₄·3H₂O (0.05 mmol) and 2.1 mg of MPA (0.09 mmol) were mixed in 4.7 mL of ultrapure water under mild stirring. Subsequently, 3 mL of NaOH solution (1.0 M) was added, causing the solution color to change from yellow to pale yellow. Then, 1 mL of NaBH₄ solution (0.1 M) was added dropwise, turning the solution yellowish-brown. After 3 hours of vigorous stirring, unreacted MPA and other impurities were removed using organic solvents and spin distillation, yielding highly pure MPA-protected Au₂₅ nanoclusters.

Preparation of the [Au₂₅(MHA)₁₈]⁻ NCs

The procedure of the [Au₂₅(MHA)₁₈]⁻ NCs is similar to the synthesis of [Au₂₅(MPA)₁₈]⁻. 4.0 mg of HAuCl₄·3H₂O (0.05 mmol) and 4.0 mL of MHA (50 mM) were mixed in 4.7 mL of ultrapure water under mild stirring. Subsequently, 0.5 mL of NaOH solution (1.0 M) was added, causing the solution color to change from yellow to pale yellow. Then, 1 mL of NaBH₄ solution (0.1 M) was added dropwise, turning the solution yellowish-brown. After 2 h of vigorous stirring, the obtained product was purified using organic solvents to remove unreacted MHA and other impurities, yielding [Au₂₅(MHA)₁₈]⁻ nanoclusters.

Preparation of the [Au₂₅(SC₆H₁₃)₁₈]⁻ NCs

HAuCl₄·3H₂O (0.125 mmol) and TOAB (0.145 mmol) were dissolved in 14 mL of THF in a 50 mL vial. After vigorous stirring for 15 minutes, the solution color changed from orange to red. Then, 1-hexanethiol (0.625 mmol) was slowly added, and the mixture was stirred for 1 hour until the red solution turned colorless. Next, freshly prepared NaBH₄ (1.25 mmol) in 2 mL of cold water was quickly added, producing bubbles as the solution turned black, indicating the formation of gold clusters. The reaction mixture was stirred for an additional 24 hours. The product was then transferred to a 25 mL round-bottom flask and dried by rotary evaporation. The dried product was redissolved in CH₂Cl₂, and the supernatant was transferred to another round-bottom flask and dried again by rotary evaporation. Finally, the product was washed with ethanol, collected by centrifugation, and the washing process was repeated at least 10 times to ensure the removal of impurities.

Electrochemical measurements

The electrochemical tests were conducted using a CHI 760E electrochemical workstation (CH Instruments Inc.) at room temperature. For HER testing, cyclic voltammetry (CV) and LSV curves were obtained in a three-electrode cell setup, consisting of a glassy carbon disk electrode (GCE, diameter 5 mm, surface area 0.07 cm²) as the working electrode, polished with alumina slurry and cleaned with ethanol and deionized water. A carbon rod served as the counter electrode, and an Ag/AgCl electrode in 0.5 M H₂SO₄ was used as the reference electrode. The potential of the reference electrode was calibrated to the reversible hydrogen electrode (RHE) using the equation: E_RHE = E_Hg/HgCl2 + (0.244 + 0.0591 × pH) V in 0.5 M H₂SO₄ solution. The working electrode for the HER was prepared by dissolving 1 mg of carbon nanotube and 1 mg of Au₂₅ NCs in 0.5 mL of CH₂Cl₂, creating an ink with 10 μL of 5 wt% Nafion. Before testing, 3.5 μL of the catalyst and 3 μL of Nafion (5 wt%) were cast onto the GCE.

The catalytic activity of Au₂₅ NCs in the CO₂RR was determined using an electrochemical workstation (CHI 600E) with a three-electrode system coupled to a CO₂ flow cell. The electrolyte solution was 1 M KOH, and the reference electrode was an Ag/AgCl electrode immersed in saturated KCl solution. An anion-exchange membrane and a platinum plate were used as the ion mobility channel and counter electrode, respectively. The working electrode was prepared by dissolving 1 mg of carbon nanotube and 1 mg of Au₂₅ NCs in 0.5 mL of CH₂Cl₂, creating a uniform dispersion ink containing 10 μL of 5 wt% Nafion solution. This solution was sprayed onto a 1 cm² gas diffusion layer (GDL) with a mass loading of 2 mg cm⁻². The potentials were converted to the RHE using the following equation:

E(RHE) = E(Ag/AgCl) + 0.197 V + 0.0591 × pH

Before the electrochemical CO₂ reduction reaction, the cathodic electrolyte was saturated with CO₂ for 30 minutes. The cathodic and anodic reaction chambers were separated using an anion exchange membrane. During the CO₂RR process, each electrolyte cell contained 30 mL of electrolyte, circulated at 40 rpm using a peristaltic pump. The gas products were analyzed quantitatively with a gas chromatograph (GC, Huaai 9560). The faradaic efficiency (FE) of the gas products was calculated using the formula:

where Q_i is the charge required to form the gaseous product, Q_total is the total charge during the reaction, N_i is the number of moles of the product detected by gas chromatography, Z is the number of electrons transferred during the formation of the product (2 for CO and H₂), and F is Faraday's constant (96 485 C mol⁻¹).

The turnover frequency (TOF) was calculated as follows:

where j_i is the partial current density of the corresponding gas product, Z is the number of transferred electrons during the formation of the product (2 for CO), F is the Faraday constant (96 485 C mol⁻¹), m_i is the mass of the loaded catalyst, ω is the relative mass fraction of Au in the catalyst, and M is the relative atomic mass of Au. Linear sweep voltammetry (LSV) was conducted in a 1 M KOH solution saturated with either N₂ or CO₂, using a scan rate of 50 mV s⁻¹.

Data availability

The data supporting this article have been included as part of the ESI.†

Author contributions

Q. T. conceived the idea. L. L. performed the theoretical calculations with help from Y. C. and F. S. X. Z. performed the experiments and analyzed the data under the guidance of L. W. L. L. wrote the manuscript with inputs from all authors. Q. T. and L. W. finalized the manuscript. All the authors approved the final version of the manuscript.

Conflicts of interest

The authors declare no competing financial interests.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (no. 22473017) and the Chongqing Science and Technology Commission (CSTB2024NSCQ-MSX0250).

References

1. Z. Zhou, Z. Pei, L. Wei, S. Zhao, X. Jian and Y. Chen, Electrocatalytic hydrogen evolution under neutral pH conditions: current understandings, recent advances, and future prospects, Energy Environ. Sci., 2020, 13, 3185–3206.
2. W. Zhai, Y. Ma, D. Chen, J. C. Ho, Z. Dai and Y. Qu, Recent progress on the long-term stability of hydrogen evolution reaction electrocatalysts, InfoMat, 2022, 4, e12357.
3. L. Xiong, Y. Qiu, X. Peng, Z. Liu and P. K. Chu, Electronic structural engineering of transition metal-based electrocatalysts for the hydrogen evolution reaction, Nano Energy, 2022, 104, 107882.
4. M. Ďurovič, J. Hnát and K. Bouzek, Electrocatalysts for the hydrogen evolution reaction in alkaline and neutral media. A comparative review, J. Power Sources, 2021, 493, 229708.
5. X. Zhang, S.-X. Guo, K. A. Gandionco, A. M. Bond and J. Zhang, Electrocatalytic carbon dioxide reduction: from fundamental principles to catalyst design, Mater. Today Adv., 2020, 7, 100074.
6. T. Ahmad, S. Liu, M. Sajid, K. Li, M. Ali, L. Liu and W. Chen, Electrochemical CO₂ reduction to C₂₊ products using Cu-based electrocatalysts: a review, Nano Res. Energy, 2022, 1, e9120021.
7. D. Xu, K. Li, B. Jia, W. Sun, W. Zhang, X. Liu and T. Ma, Electrocatalytic CO₂ reduction towards industrial applications, Carbon Energy, 2023, 5, e230.
8. X. An, S. Li, X. Hao, Z. Xie, X. Du, Z. Wang, X. Hao, A. Abudula and G. Guan, Common strategies for improving the performances of tin and bismuth-based catalysts in the electrocatalytic reduction of CO₂ to formic acid/formate, Renewable Sustainable Energy Rev., 2021, 143, 110952.
9. X. Mao, W. Gong, Y. Fu, J. Li, X. Wang, A. P. O'Mullane, Y. Xiong and A. Du, Computational Design and Experimental Validation of Enzyme Mimicking Cu-Based Metal–Organic Frameworks for the Reduction of CO₂ into C₂ Products: C–C Coupling Promoted by Ligand Modulation and the Optimal Cu–Cu Distance, J. Am. Chem. Soc., 2023, 145, 21442–21453.
10. K. Sokołowska, S. Malola, M. Lahtinen, V. Saarnio, P. Permi, K. Koskinen, M. Jalasvuori, H. Häkkinen, L. Lehtovaara and T. Lahtinen, Towards controlled synthesis of water-soluble gold nanoclusters: synthesis and analysis, J. Phys. Chem. C, 2019, 123, 2602–2612.
11. P. Maity, S. Xie, M. Yamauchi and T. Tsukuda, Stabilized gold clusters: from isolation toward controlled synthesis, Nanoscale, 2012, 4, 4027–4037.
12. K. Kwak and D. Lee, Electrochemistry of atomically precise metal nanoclusters, Acc. Chem. Res., 2018, 52, 12–22.
13. Y. Jo, M. Choi, M. Kim, J. S. Yoo, W. Choi and D. Lee, Promotion of alkaline hydrogen production via Ni-doping of atomically precise Ag nanoclusters, Bull. Korean Chem. Soc., 2021, 42, 1672–1677.
14. S. Zhao, R. Jin and R. Jin, Opportunities and challenges in CO₂ reduction by gold-and silver-based electrocatalysts: from bulk metals to nanoparticles and atomically precise nanoclusters, ACS Energy Lett., 2018, 3, 452–462.
15. Q. Tang, G. Hu, V. Fung and D.-e. Jiang, Insights into interfaces, stability, electronic properties, and catalytic activities of atomically precise metal nanoclusters from first principles, Acc. Chem. Res., 2018, 51, 2793–2802.
16. F. Sun, C. Deng, S. Tian and Q. Tang, Oxygen electrocatalysis by [Au₂₅(SR)₁₈]: charge, doping, and ligand removal effect, ACS Catal., 2021, 11, 7957–7969.
17. F. Sun, Q. Tang and D.-e. Jiang, Theoretical advances in understanding and designing the active sites for hydrogen evolution reaction, ACS Catal., 2022, 12, 8404–8433.
18. T. Kawawaki, T. Okada, D. Hirayama and Y. Negishi, Atomically precise metal nanoclusters as catalysts for electrocatalytic CO₂ reduction, Green Chem., 2024, 26, 122–163.
19. W. Choi, H. Seong, V. Efremov, Y. Lee, S. Im, D.-H. Lim, J. S. Yoo and D. Lee, Controlled syngas production by electrocatalytic CO₂ reduction on formulated Au₂₅(SR)₁₈ and PtAu₂₄(SR)₁₈ nanoclusters, J. Chem. Phys., 2021, 155, 014305.
20. D. R. Alfonso, D. Kauffman and C. Matranga, Active sites of ligand-protected Au₂₅ nanoparticle catalysts for CO₂ electroreduction to CO, J. Chem. Phys., 2016, 144, 184705.
21. H. Seong, V. Efremov, G. Park, H. Kim, J. S. Yoo and D. Lee, Atomically precise gold nanoclusters as model catalysts for identifying active sites for electroreduction of CO₂, Angew. Chem., 2021, 133, 14684–14691.
22. S. M. Han, M. Park, J. Kim and D. Lee, Boosting the Electroreduction of CO₂ to CO by Ligand Engineering of Gold Nanoclusters, Angew. Chem., 2024, 63, e202404387.
23. O. López-Estrada, N. Mammen, L. Laverdure, M. M. Melander, H. Häkkinen and K. Honkala, Computational Criteria for Hydrogen Evolution Activity on Ligand-Protected Au₂₅-Based Nanoclusters, ACS Catal., 2023, 13, 8997–9006.
24. Y. Li, S. Li, A. V. Nagarajan, Z. Liu, S. Nevins, Y. Song, G. Mpourmpakis and R. Jin, Hydrogen evolution electrocatalyst design: turning inert gold into active catalyst by atomically precise nanochemistry, J. Am. Chem. Soc., 2021, 143, 11102–11108.
25. K. Kwak, W. Choi, Q. Tang, D.-e. Jiang and D. Lee, Rationally designed metal nanocluster for electrocatalytic hydrogen production from water, J. Mater. Chem. A, 2018, 6, 19495–19501.
26. G. Hu, Q. Tang, D. Lee, Z. Wu and D.-e. Jiang, Metallic hydrogen in atomically precise gold nanoclusters, Chem. Mater., 2017, 29, 4840–4847.
27. F. Sun, L. Qin, Z. Tang and Q. Tang, Revisiting the activity origin of the PtAu₂₄(SR)₁₈ nanocluster for enhanced electrocatalytic hydrogen evolution by combining first-principles simulations with the experimental in situ FTIR technique, Chem. Sci., 2024, 15, 16142–16155.
28. D. R. Kauffman, D. Alfonso, C. Matranga, H. Qian and R. Jin, Experimental and computational investigation of Au₂₅ clusters and CO₂: a unique interaction and enhanced electrocatalytic activity, J. Am. Chem. Soc., 2012, 134, 10237–10243.
29. K. Kwak, W. Choi, Q. Tang, M. Kim, Y. Lee, D.-e. Jiang and D. Lee, A molecule-like PtAu₂₄(SC₆H₁₃)₁₈ nanocluster as an electrocatalyst for hydrogen production, Nat. Commun., 2017, 8, 14723.
30. E. Pensa, L. M. Azofra, R. C. Salvarezza and P. Carro, Effect of ligands on the stability of gold nanoclusters, J. Phys. Chem. Lett., 2022, 13, 6475–6480.
31. S. Li, A. V. Nagarajan, Y. Li, D. R. Kauffman, G. Mpourmpakis and R. Jin, The role of ligands in atomically precise nanocluster-catalyzed CO₂ electrochemical reduction, Nanoscale, 2021, 13, 2333–2337.
32. B. Kumar, T. Kawawaki, N. Shimizu, Y. Imai, D. Suzuki, S. Hossain, L. V. Nair and Y. Negishi, Gold nanoclusters as electrocatalysts: size, ligands, heteroatom doping, and charge dependences, Nanoscale, 2020, 12, 9969–9979.
33. S. F. Yuan, R. L. He, X. S. Han, J. Q. Wang, Z. J. Guan and Q. M. Wang, Robust gold nanocluster protected with amidinates for electrocatalytic CO₂ reduction, Angew. Chem., Int. Ed., 2021, 60, 14345–14349.
34. J. Wang, F. Xu, Z. Y. Wang, S. Q. Zang and T. C. Mak, Ligand-Shell Engineering of a Au₂₈ Nanocluster Boosts Electrocatalytic CO₂ Reduction, Angew. Chem., Int. Ed., 2022, 61, e202207492.
35. X. Li, S. Takano and T. Tsukuda, Ligand effects on the hydrogen evolution reaction catalyzed by Au₁₃ and Pt@Au₁₂: alkynyl vs. thiolate, J. Phys. Chem. C, 2021, 125, 23226–23230.
36. J. C. Love, L. A. Estroff, J. K. Kriebel, R. G. Nuzzo and G. M. Whitesides, Self-assembled monolayers of thiolates on metals as a form of nanotechnology, Chem. Rev., 2005, 105, 1103–1170.
37. C. Vericat, M. Vela, G. Benitez, P. Carro and R. Salvarezza, Self-assembled monolayers of thiols and dithiols on gold: new challenges for a well-known system, Chem. Soc. Rev., 2010, 39, 1805–1834.
38. R. G. Nuzzo and D. L. Allara, Adsorption of bifunctional organic disulfides on gold surfaces, J. Am. Chem. Soc., 1983, 105, 4481–4483.
39. Z. Liu, H. Tan, B. Li, Z. Hu, D.-e. Jiang, Q. Yao, L. Wang and J. Xie, Ligand effect on switching the rate-determining step of water oxidation in atomically precise metal nanoclusters, Nat. Commun., 2023, 14, 3374.
40. S. Yoo, S. Yoo, G. Deng, F. Sun, K. Lee, H. Jang, C. W. Lee, X. Liu, J. Jang, Q. Tang, Y. J. Hwang, T. Hyeon and M. S. Bootharaju, Nanocluster surface microenvironment modulates electrocatalytic CO₂ reduction, Adv. Mater., 2024, 36, 2313032.
41. S. Zhao, N. Austin, M. Li, Y. Song, S. D. House, S. Bernhard, J. C. Yang, G. Mpourmpakis and R. Jin, Influence of atomic-level morphology on catalysis: the case of sphere and rod-like gold nanoclusters for CO₂ electroreduction, ACS Catal., 2018, 8, 4996–5001.
42. N. Austin, S. Zhao, J. R. McKone, R. Jin and G. Mpourmpakis, Elucidating the active sites for CO₂ electroreduction on ligand-protected Au₂₅ nanoclusters, Catal. Sci. Technol., 2018, 8, 3795–3805.
43. S. Li, D. Alfonso, A. V. Nagarajan, S. D. House, J. C. Yang, D. R. Kauffman, G. Mpourmpakis and R. Jin, Monopalladium substitution in gold nanoclusters enhances CO₂ electroreduction activity and selectivity, ACS Catal., 2020, 10, 12011–12016.
44. S. Li, A. V. Nagarajan, D. R. Alfonso, M. Sun, D. R. Kauffman, G. Mpourmpakis and R. Jin, Boosting CO₂ electrochemical reduction with atomically precise surface modification on gold nanoclusters, Angew. Chem., Int. Ed., 2021, 60, 6351–6356.
45. R. Jin, Quantum sized, thiolate-protected gold nanoclusters, Nanoscale, 2010, 2, 343–362.
46. P. Hu, L. Chen, X. Kang and S. Chen, Surface functionalization of metal nanoparticles by conjugated metal–ligand interfacial bonds: impacts on intraparticle charge transfer, Acc. Chem. Res., 2016, 49, 2251–2260.
47. F. Sun, L. Qin, Z. Tang, G. Deng, M. S. Bootharaju, Z. Wei, Q. Tang and T. Hyeon, SR removal or –R removal? A mechanistic revisit on the puzzle of ligand etching of Au₂₅(SR)₁₈ nanoclusters during electrocatalysis, Chem. Sci., 2023, 14, 10532–10546.
48. T. Giorgino, Computing 1-D atomic densities in macromolecular simulations: the density profile tool for VMD, Comput. Phys. Commun., 2014, 185, 317–322.
49. X. Zhao and Y. Liu, Origin of selective production of hydrogen peroxide by electrochemical oxygen reduction, J. Am. Chem. Soc., 2021, 143, 9423–9428.
50. J. VandeVondele, M. Krack, F. Mohamed, M. Parrinello, T. Chassaing and J. Hutter, Quickstep: fast and accurate density functional calculations using a mixed Gaussian and plane waves approach, Comput. Phys. Commun., 2005, 167, 103–128.
51. J. P. Perdew, K. Burke and M. Ernzerhof, Generalized gradient approximation made simple, Phys. Rev. Lett., 1996, 77, 3865.
52. W. G. Hoover, Canonical dynamics: equilibrium phase-space distributions, Phys. Rev. A: At., Mol., Opt. Phys., 1985, 31, 1695.
53. G. J. Martyna, M. L. Klein and M. Tuckerman, Nosé–Hoover chains: the canonical ensemble via continuous dynamics, J. Chem. Phys., 1992, 97, 2635–2643.
54. S. Grimme, J. Antony, S. Ehrlich and H. Krieg, A consistent and accurate ab initio parametrization of density functional dispersion correction (DFT-D) for the 94 elements H-Pu, J. Chem. Phys., 2010, 132, 154104.
55. S. Grimme, S. Ehrlich and L. Goerigk, Effect of the damping function in dispersion corrected density functional theory, J. Comput. Chem., 2011, 32, 1456–1465.
56. J. VandeVondele and J. Hutter, Gaussian basis sets for accurate calculations on molecular systems in gas and condensed phases, J. Chem. Phys., 2007, 127, 114105.
57. G. Kresse and J. Furthmüller, Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 11169.
58. P. E. Blöchl, Projector augmented-wave method, Phys. Rev. B: Condens. Matter Mater. Phys., 1994, 50, 17953.
59. I. L. Garzon and A. Posada-Amarillas, Structural and vibrational analysis of amorphous Au₅₅ clusters, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 11796.
60. R. W. Hall and B. J. Berne, Nonergodicity in path integral molecular dynamics, J. Chem. Phys., 1984, 81, 3641–3643.
61. M. Sprik and G. Ciccotti, Free energy from constrained molecular dynamics, J. Chem. Phys., 1998, 109, 7737–7744.
62. A. A. Hassanali, J. Cuny, V. Verdolino and M. Parrinello, Aqueous solutions: state of the art in ab initio molecular dynamics, Philos. Trans. R. Soc., A, 2014, 372, 20120482.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4sc07181f

‡ These two authors contributed equally to this work.

---